The experiment was conducted at Laizhou Shunchang Fisheries Co., Ltd. in China. Wild adult black scrapers between 2 and 3 years old were collected from the Yellow Sea and reared in incubation ponds (length × width × height, 6.0 × 5.0 × 1.0 m, respectively). The broodstock naturally spawned and fertilised in the pool after implementing reproductive control measures, such as artificial domestication, water temperature regulation, and nutrient enhancement. The fertilised eggs (spherical and viscous) were collected and incubated for 40–44 h under a salinity of 32–33, pH 7.0–8.0, and continuous micro-inflation. The water temperature was 22.0 ± 0.5°C for the rearing of early larvae, after which the temperature was gradually raised to 24.0 ± 0.5°C for the rearing of the juveniles. Seawater was subjected to secondary sand filtration during the larval rearing. The rearing pool was maintained under natural light during the day, and the lights were turned off at night. After hatching, the larvae were fed a series of different feeds including SS-type Brachionus plicatilis (rotifers) (5–11 days post-hatching, [dph]), L-type rotifer (9–22 dph), Artemia nauplii (19–32 dph), A. adults (30–37 dph), and artificial compound feed (32–40 dph) as their growth progressed. The wet weight of the rotifer was used in this experiment.

All data were analysed using SPSS 23.0 software (IBM, Armonk, NY, USA). The diagram was plotted using Excel or OriginPro 9.1 software (OriginLab, Northampton, MA, USA). All data are reported as the means ± standard deviation (SD). The normality and homoscedasticity of the continuous data was tested using Shapiro–Wilk tests. Differences between independent samples were assessed using a one-way analysis of variance with Tukey’s Honestly Significant Difference (HSD) test. All results with P < 0.05 were considered statistically significant.

Morphological changes during the development of black scrapers are shown in Figure 2 . The characteristics of the larvae, juveniles, and adult fish during each developmental period are described below.

The growth rate (assessed as change in body weight [BW] over time) was slow before 20 dph, increased from 20–40 dph, and increased again from 40–60 dph ( Figure 3A ). Furthermore, the growth rate of body length (BL) pre-anal length (PL), and snout length (SL) was slow before 20 dph, accelerated from 20–36 or 38 dph, and increased again from 36 or 38–60 dph ( Figures 3B–D ). Similarly, the growth rate of body height (BH) and head length (HL) increased at 18 or 24 dph and increased again from 40–60 dph ( Figures 3E, F ). The total length (TL) and eye orbit diameter (ED) accelerated at 18 or 20 dph ( Figures 3G, H ). Therefore, we speculated that there are two growth turning points for BW, BL, PL, SL, BH, and HL, with the first turning point occurring at 18, 20, or 24 dph and the second at 36, 38, or 40 dph. Furthermore, we speculated that the two turning points appeared at 18 or 20 dph for TL and ED.

ACP levels were significantly higher at 4 and 25 dph than at other ages except at 20 dph (P< 0.05) and were significantly higher at 20 dph than at 0, 8, 12, 30, and 35 dph (P < 0.05; Figure 4A ). AKP levels were significantly higher at 8, 12, and 16 dph than at the other ages (P < 0.05, Figure 4B ). ACPP levels at 0 and 4 dph and AKPP levels at 4, 20, and 25 dph were significantly higher than those at other ages (P < 0.05; Figures 4C, D ). Amylase levels were significantly higher at 4, 25, 30, and 35 dph than that at other ages (P < 0.05, Figure 4E ). Lipase levels were significantly higher at 4 and 16 dph than at 8, 12, 30, and 35 dph (P < 0.05, Figure 4F ).

The T3 level gradually significantly increased from 0 to 8 dph, with a plateau at 12–30 dph, and then significantly increased at 35 dph (P < 0.05, Figure 5A ). T4 levels gradually increased from 0–25 dph and increased significantly at 25, 30, and 35 dph compared with that at other ages (P < 0.05, Figure 5B ). GH levels increased from 0–8 dph, with a plateau at 8–20 dph, and then increased again for 25–30–35 dph ( Figure 5C ). IGF-1 levels increased considerably with a plateau between 8–20 dph and then decreased considerably at 30 and 35 dph ( Figure 5D ).

No differences were observed in the activities of the six digestive enzymes between normal and abnormal fish at 30 dph ( Figure 6A ). However, T4 and IGF-1 levels in normal fish were significantly higher than those in abnormal fish (P < 0.05, Figure 6B ).

Previous studies showed that the growth rates of teleosts vary at different developmental stages, indicating that stages characterise the growth rate and are primarily expressed by growth turning points in the early stages, such as in Limanda yokohamae (Günther, 1877) (Fukuhara, 1988; Hopkins, 1992; Houde, 1997). In the thick-lipped grey mullet, Chelon labrosus, morphometric growth distinguished three distinct developmental stages separated by two morphometric lengths, with the first inflection point at 14 dph and the second at 25 dph (Ben Khemis et al., 2013). In the present study, there were two turning points in growth, as indicated by BW and TL values, from 0–60 dph. The first turning point appeared at 20 dph as showed by BW, BL, PL and SL. Similar results were observed by Guan et al. (2013), who also observed a growth turning point for the TL of T. modestus at 20 dph. This phenomenon may be related to changes in the morphology and diet of larval fish, with a more efficient digestive system formed at this stage (Guan et al., 2013). The fish calibre increased, and the digestive system gradually developed after 20 dph. The fry could swallow and digest the A. nauplii; thus, nutrition increased, and their growth rate accelerated. The second turning point appeared at 40 dph for BW, BH, and HL, and at 36 dph for PL and SL; this may be due to the consumption of compound feed at approximately 36 dph, as it contains more abundant nutrients than A. nauplii and is suitable for rapid growth.

With the formation of digestive organs and glands in the early developmental stage of fish, the digestive enzymes secreted in the body undergo specific changes according various stages of growth (Buddington, 1985; Govoni et al., 1986). Thus, changes in digestive enzyme activity in the body may be caused by individual growth and digestive organ development (Comabella et al., 2006), but also changes in feed types (Morais et al., 2004), and changes in enzyme expression caused by juvenile metabolism (Alvarez-González et al., 2008). In the present study, the two critical periods for quantitative changes in digestive enzyme activity were the transition period from endogenous to exogenous nutrition in fry and the transition period from fry to juveniles (Cousin et al., 1987). Furthermore, the activities of ACP, AKP, ACPP, AKPP, amylase, and lipase were detected in the newly hatched larvae, as reported in the common snook, Centropomus undecimalis, with all digestive enzymatic activities detected during larviculture (Jimenez-Martinez et al., 2012), suggesting that the larvae can digest food when it transitions from endogenous to exogenous nutrition. In the red drum, Sciaenops ocellatus, digestive enzymes for proteins, lipids and amylase were present in the larvae, and their activity subsequently increased with age and length (Lazo et al., 2007). In our study, the larvae start feeding on the third day after hatching; therefore, juvenile digestive enzyme activity may be related to gene expression from endogenous nutrition, not food intake.

In the present study, ACP levels were much lower than those of AKP from 0–20 dph. Similar results were obtained for seabass, Lates calcarifer (Walford and Lam, 1993). For the AKP levels, they significantly increased from 0–12 dph, peaked at 12 dph, and significantly decreased at 16 dph. Similar results have been reported for Eurasian perch, Perca fluviatilis (Cuvier-Péres and Kestemont, 2001), Persian sturgeon, Acipenser persicus (Babaei et al., 2011), and sea bass, Dicentrarchus labrax (Infante and Cahu, 1994). These results suggest that AKP activity first increased and then decreased to finally reach a stable state, a common feature during the individual development of vertebrates and an essential step in the process of metamorphosis. Concerning ACP levels, they significantly increased at 4 dph in T. modestus. This increase may be due to the fact that larvae were fed on rotifers, as it was shown that the ACP activity in the bodies of folded armyworm rotifers was five times higher than that of AKP (Wethmar and Kleinow, 1993); and thus they consumed exogenous enzymes. In the common carp, Cyprinus carpio, increased AKP activity can be related to the adaptation of larvae to digest protein content in the food (Farhoudi et al., 2013). Most studies suggest that as gastric function gradually develops and improves, ACP activity increases, and AKP activity generally decreases (Ribeiro et al., 1999; Comabella et al., 2006; Alvarez-González et al., 2008). These results are similar to those observed in our study. From 12–25 dph, the ACP and AKP levels increased and decreased, respectively, which may be because the stomach gradually develops at approximately 12 dph, the gastric glands may appear and it is known that protein digestion mainly depends on acidic digestion in the stomach. However, the specific time at which gastric glands appear in T. modestus requires further study.

In the present study, ACPP activity significantly increased from hatching. Both acid and alkaline phosphatase activities have been detected during larval development in different marine fishes (Cahu and Infante, 1994; Moyano et al., 1996). In California halibut, Paralichthys californicus, acid phosphatases have been detected in the early stages of development, although ACPP levels were higher at hatching and during yolk sac absorption (Alvarez-González et al., 2005). Once yolk-sac reserves were depleted, acid phosphatase activity decreased below alkaline phosphatase activity levels in this species. In white seabream, Diplodus sargus, ACPP-specific activity increased in this species throughout larval development, peaking at 20 dph (Guerreiro et al., 2010). Similar changes were observed in the senegal sole (Martinez et al., 1999) and bay snook, Petenia splendida (Uscanga-Martínez et al., 2011); this may be because the hatched larvae directly contact the external environment and lose the protection of the egg membrane; thus, higher ACPP activity is required to regulate physiological metabolism and immune protection. Indeed, intestinal AKPP are involved in the absorption of nutrients, such as lipids, glucose, calcium, and inorganic phosphate (Roubaty and Portmann, 1988; Dupuis et al., 1991; Mahmood et al., 1994; Tengjaroenkul et al., 2000). In Senegal sole, AKPP activity decreased from 5–20 dph but subsequently showed a sharp increase, possibly linked to the development of enterocytes (Martinez et al., 1999). In California halibut, AKPP was detected at maximal activity levels at 4 dph, decreasing after a few days and increasing again until the second maximum activity at 15 dph (Alvarez-González et al., 2005). In the present study in T. modestus, AKPP activity significantly increased at 4 dph, and ACPP activity increased significantly at 0 and 4 dph, indicating the significance of intestinal absorption of these two phosphatases. A sharp increase in AKPP activity was detected at 20 and 25 dph, which may reflect the development of the brush border membrane of enterocytes and, hence, an increase in the relevance of the absorptive processes mediated by these enzymes in T. modestus.

In the present study, amylase levels were significantly higher at 4 dph then decreased until 20 dph. A previous study in turbot, Scophthalmus maximus, the authors proposed that this may be due to the consumption of Artemia; therefore, a change in the feeding pattern should result in a significant reduction in amylase activity (Cousin et al., 1987). At 4 dph, the transition period from endogenous to exogenous nutrition suggests that the black scraper could absorb carbohydrates after the full development of the mouth opening; then, it gradually decreased as the larvae developed until 20 dph. However, in seabass larvae, the initial high levels of amylase at 4 dph did not seem to be food induced (enriched rotifers), and may be better explained by programmed gene expression (Infante and Cahu, 2001). Furthermore, we found that amylase levels were significantly higher at 25, 30, and 35 dph, the transition period from Artemia as the food source to compound feed. In the leopard grouper, Mycteroperca rosacea, amylase levels increased significantly between 30 and 40 dph, suggesting that larvae could be successfully weaned from micro-diets during this period (Martínez-Lagos et al., 2014). In the present study, the increase in amylase activity at 30 and 35 dph was likely influenced by the carbohydrate content of the administered compound diet, which contained carbohydrates derived from yeast extract included in the formulation. Similar amylase patterns, reported in other studies, are thought to be influenced by the carbohydrates in compound diets (Cara et al., 2003; Zouiten et al., 2008).

Lipase activity was detected throughout the early developmental stages in our study, and similar results were found for California halibut (Alvarez-González et al., 2005) and common dentex, Dentex dentex (Gisbert et al., 2009). Furthermore, it was previously reported that lipase was actively found in 4 dph T. modestus larvae (Gwak and Lee, 2009). However, lipase activity was detected until 13 dph in the Mayan cichlid, Cichlasoma urophthalmus (López-Ramírez et al., 2011) and until 20 dph in turbot (Cousin et al., 1987). In spotted sand bass, lipase activity was detected immediately after mouth opening, gradually increased until 15 dph, and remained stable until the end of the study period (Alvarez-González et al., 2008). In Japanese flounder, P. olivaceus, neutral lipase activity showed substantial fluctuations in larvae during the first few days (Bolasina et al., 2006). In the present study, lipase activity was significantly higher at 4 and 16 dph. Studies showed that early larvae contain two types of lipases, phosphatase A2 being involved in the digestion of fat in the yolk sac, whereas lipase is involved in fat digestion by the intestinal epithelium (Oozeki and Bailey, 1995; Izquierdo et al., 2000). In the present study, the unstable changes in lipase activity after the first feeding of juveniles may have been caused by the mutual conversion of the two types of lipases, while the decrease in lipase activity in the later stage may have been due to the rich protein and low-fat content in the feed. In the Miiuy croaker (Miichthys miiuy) larvae, amylase and lipase activities can be modulated by feeding modes (Shan et al., 2009). In addition, the pattern of digestive enzyme activity was related to organogenesis and the type of food used at different developmental stages in blackspot seabream, Pagellus bogaraveo (Ribeiro et al., 2008). Therefore, in the present study, the changes in these enzyme activities may be related to the type of bait, the digestive system’s degree of development and the different growth stages. If we understand the changes in digestive enzyme activities during growth and development, it can be used as a reference for improving the weaning process for the larviculture of T. modestus in the future.

The ontogeny of thyroid hormones is related to specific morphological changes such as fins, eyes, and stomach, which characterized of early development in the gilthead seabream, Sparus aurata (Szisch et al., 2005). Significant levels of T3 and T4 were found in the eggs of 26 species of various freshwater, marine, and diadromous teleosts, suggesting that thyroid hormones play crucial roles in egg development (Tagawa et al., 1990b). In the present study, T3 and T4 were detected at relatively low levels in the newly hatched larvae of T. modestus. Similar results have been observed for Japanese flounder (Miwa et al., 1988) and black rockfish, Sebastes schlegelii (Chin et al., 2010). Furthermore, de Jesus et al. (1991) detected T3 and T4 in newly fertilised eggs of the Japanese flounder, suggesting a synergistic action of T3 and T4 during metamorphosis in the flounder. It is reported that the higher concentrations of T3 than T4 and the sharp decrease in T3 concentration immediately after hatching indicates a primary role of T3 in the early development of the flounder embryos and larvae (Tagawa et al., 1990a). In the present study, the low concentration of thyroid hormones in newly hatched larvae may be related to the low level of thyroid development and the consumption or short maintenance time of thyroid hormones (primarily maternal) during embryonic development. Furthermore, the T3 and T4 levels showed an increasing trend with growth and development. Similar results were reported in Spotted halibut, Verasper variegatus, and T4 levels increased gradually during the early metamorphic stage, reaching peak levels at the metamorphosis climax (Hotta, 2001). In silver sea bream Sparus sarba larvae, T3 and T4 were detected at 1 dph, and both increased as development progressed with a distinct increase between 21 and 35 dph (Deane and Woo, 2003b). These results are consistent with those of the present study. In our study, T4 levels were higher than that of T3 at every stage. Similar results have been observed in rabbitfish, Siganus guttatus (Ayson and Lam, 1993), spotted grouper, Epinephelus tauvina (Leatherland et al., 1990), and Atlantic halibut (Einarsdóttir et al., 2006); this may be due to the presence of 5’-monodeiodinase in the body, which converts a portion of T4 into T3 through deiodination in the peripheral tissues. However, there is insufficient evidence to confirm this hypothesis here.

GH and IGF-I play important roles as modulators of development, growth, and reproduction in teleosts (Funes et al., 2006). A study in orange-spotted grouper E. coioides showed that gh mRNA expression was detected in 1 dph larvae, and gh present in eggs and larvae may play a vital role in the early development, especially during the metamorphosis of its larvae (Li et al., 2005). Similarly, transcripts for IGF-I were also detected throughout development in unfertilised eggs, embryos, and larvae in gilthead seabream, suggesting that IGF-I may play a role during the early development of teleosts (Perrot et al., 1999). In the present study, the GH levels gradually increased from 0–8 dph, with a plateau at 8–20 dph, and then a second increase for 25–30–35 dph; and IGF-1 levels showed a large increase with a plateau observed between 8–20 dph and then a large decrease at 30 and 35 dph. Similarly, GH abundance reached a constant and high level in the silver sea bream from 35 dph onward. In contrast, IGF-I level peaked at 35 dph and then significantly decreased (Deane et al., 2003a). The highest amounts of gh mRNA were observed at 25 and 50 dph in Siberian sturgeon, Acipenser baerii larvae (Abdolahnejad et al., 2015). Furthermore, the IGF-1 level was much higher than that of GH during development in the present study. These results suggest that the function of GH in promoting fish growth is closely related to IGF-1, and that the two hormones are synchronised in promoting the growth and development of fish. However, differently, GH and IGF-1 mRNA levels showed approximately inverse expression patterns in the thick-lipped grey mullet (Gilannejad et al., 2020), suggesting that the changes of GH and IGF-1 may be species-specific. In the present study, IGF-1 at 25 dph and GH at 30 dph were significantly decreased, possibly because the juveniles were in a critical metamorphosis stage. Study also shows that fish secrete more thyroid hormones during metamorphosis to adapt to this physiological change and reduce the synthesis and secretion of GH (Shiao and Hwang, 2006; Campinho, 2019).

Many studies showed that digestive enzyme activity and hormone content change during the transition from larvae to juveniles (metamorphosis) (Blaxter, 1988). In senegal sole, the activities of ACPP, AKPP, and other digestive enzymes decreased during metamorphosis (Martinez et al., 1999). In the present study, digestive enzyme activity in abnormal T. modestus fish was lower than in normal fish, suggesting that most digestive enzymes have low activity during metamorphosis. However, they play an indispensable role in the physiological functions of metamorphosis.

Several studies have shown that the thyroid-growth hormone axis plays an important role in regulating metamorphosis in fish. In orange-spotted grouper, E. coioides, a higher dose of thyroid hormones was appropriate for acceleration of metamorphosis and improvement of survival in 3- and 4-week-old larvae (de Jesus et al., 1998). Furthermore, exogenous thyroid hormone induces premature differentiation of the pectoral fins and accelerates the growth of the pelvic fins in zebrafish (Brown, 1997). In the present study, the levels of T4 and IGF-1 in abnormal fish were significantly lower than those in normal fish. Similar results were observed in Atlantic halibut. Between normally and abnormally metamorphosing larvae, IGF-I content was reduced in the abnormal fish, indicating that IGF-I either has a regulatory role in metamorphosis, or is affected by abnormal metamorphosis (Hildahl et al., 2007). In tongue sole, Cynoglossus semilaevis, IGF-1 plays an important role in the physiological functions related to metamorphosis (Duan, 1998). These results suggest that T4 and IGF-1 play important roles in the normal development of T. modestus.